Associations of Child Maltreatment and Intimate Partner Violence With Elder Abuse in a US Chinese Population

Key Points

Question  Are adults with a history of child maltreatment and/or intimate partner violence at higher odds of experiencing elder abuse?

Findings  In a cross-sectional study of 3157 US Chinese older adults in Chicago, Illinois, individuals with a history of child maltreatment had 2 times higher odds of experiencing intimate partner violence and elder abuse, and those who had experienced intimate partner violence had 6 times higher odds of experiencing elder abuse.

Meaning  Health care professionals should be more aware of the possibility of abuse when a patient has a known history of exposure to violence and consider the cumulative effect of violence among those experiencing elder abuse.

Importance  People who have experienced abuse as a child or violence with an intimate partner might have higher odds of being abused again, but this has been insufficiently investigated regarding elder abuse. More conclusive evidence might be critical to assessment and prevention strategies.

Objective  To examine the associations of child maltreatment and intimate partner violence with elder abuse.

Design, Setting, and Participants  Cross-sectional data of 3157 community-dwelling US Chinese older adults (60 years or older) in Chicago, Illinois, were collected during 2011 through 2013.

Exposures  Cases of child maltreatment and intimate partner violence.

Main Outcomes and Measures  Cases of elder abuse.

Results  Of the 3157 US Chinese older adults included in the study, 1328 (42.1%) were men, and the mean (SD) age was 72.8 (8.3) years. The prevalence of elder abuse, child maltreatment, and intimate partner violence in the cohort was 15.2%, 11.4%, and 6.5%, respectively. After adjusting for sociodemographic variables, health status, quality of life, and health change, individuals reporting child maltreatment had increased odds of intimate partner violence (13.4% vs 5.6%; adjusted odds ratio [aOR], 2.57; 95% CI, 1.78-3.71) and elder abuse (25.2% vs 13.8%; aOR, 2.08; 95% CI, 1.57-2.75) than those not reporting child maltreatment. Individuals reporting intimate partner violence had increased odds of elder abuse than those not reporting intimate partner violence (48.8% vs 12.9%; aOR, 5.53; 95% CI, 4.01-7.64).

Conclusions and Relevance  Prior abuse across major lifespan stages is associated with higher odds of elder abuse. Health care professionals should be more aware of the possibility of abuse when there is a known history of violence in a patient and consider the cumulative effect of violence among those exposed to elder abuse.

Interpersonal violence refers to violence between individuals and is subdivided into family, intimate partner, acquaintance, and stranger violence.¹ It is a pervasive public health issue associated with increased morbidity, mortality, and social and health care costs.^2-10 Research suggests that people exposed to one form of violence might be at higher risks of experiencing other forms of violence.^11-13 Within each age-related stage, the World Health Organization has identified child maltreatment, intimate partner violence, and elder abuse as common types of interpersonal violence.¹ After age 60 years, approximately 11% of people in the United States have experienced elder abuse, and the rates are higher among minority populations.^3,14,15 Elder abuse has been linked to significant health-related consequences.^3,16-18

Identifying individuals with increased risk of abuse among older populations is vital for elder abuse prevention and intervention. There exists a substantial body of literature that suggests a history of sexual abuse as a significant indicator for exposure to subsequent abuse among children and women.^11,19-22 Although this is important, there is limited information about different forms of abuse across major lifespan stages and their relationships to elder abuse.

Building on the previous work, this study seeks to bridge the knowledge gap by examining lifespan abuse (ie, psychological, physical, sexual) and experience of elder abuse in a large US Chinese older adult cohort. The inquiry was addressed by 3 specific relationships: (1) child maltreatment with intimate partner violence, (2) child maltreatment with elder abuse, and (3) intimate partner violence with elder abuse. The primary hypothesis is that prior abuse is associated with increased odds of being abused again.

The present study used the baseline data (from 2011-2013) of the Population Study of Chinese Elderly in Chicago project. Within a sample (3157 individuals) that is representative of the Chinese aging population in the greater Chicago area, the Population Study of Chinese Elderly in Chicago study aims to examine the key determinants of health and well-being of the population.²³ The eligibility criteria were (1) community-dwelling older adults 60 years and older and (2) older adults who self-identified as Chinese. With informed consent, the participants were interviewed by multilingual research assistants in Cantonese, Toishanese, Mandarin, English, and Teochew in face-to-face and private context. Data were collected using a custom-built, web-based software that recorded English and Chinese (traditional/simplified) simultaneously.²⁴ This study was approved by the institutional review board of the Rush University Medical Center.

Elder abuse prevalence (after age 60 years) was assessed by a self-reported measure modified from the Hwalek-Sengstok Elder Abuse Screening Test and the Vulnerability to Abuse Screening Scale.^25,26 The measure contains 10 questions asking indicative information related to elder abuse (eg, “Are you afraid of anyone in your family?”; “Do you feel nobody wants you around?”; “Does anyone tell you that you give them too much trouble?”; “Has anyone forced you to do things you didn’t want to do?”). Affirmative answers to any item were considered as elder abuse cases. The 10-item measure has shown adequate reliability (Cronbach α = 0.80), content validity, and convergent validity among Chinese aging populations in the United States and China.^27-30

Child maltreatment (before age 18 years) and intimate partner violence (within age 18-59 years) experience in physical, psychological, and sexual aspects were measured respectively by a 5-item Extended Hurt, Insult, Threaten, Scream scale.^31,32 Participants were asked whether they had the experience of being (1) physically hurt, (2) insulted, (3) threatened with harm, (4) screamed and cursed at, and (5) touched intimately when it was unwanted. Individuals with affirmative responses to any item were regarded as child maltreatment or intimate partner violence cases. The Extended Hurt, Insult, Threaten, Scream scale has been used to assess intimate partner violence among a Chinese population and showed good internal consistency, concurrent, and discriminant validity.³² The mean interitem correlation is 0.18 for child maltreatment and 0.37 for intimate partner violence in the population.

Basic sociodemographic information was collected including age (years), gender (male and female), education (years of education completed), personal annual income ($0-$4999, $5000-$9999, $10 000-$14 999, $15 000-$19 999, $20 000-$24 999, $25 000-$29 999, $30 000-$34 999, $35 000-$49 999, $50 000-$74 999, or ≥$75 000), marital status (married or not married), living arrangement (number of people lived with), years in the United States, number of children, country of origin (mainland China, Hong Kong, Taiwan, United States/Canada, or other), and language preference (Cantonese/Toishanese or Mandarin/English). Overall health status, quality of life, and health change were assessed by asking (1) “In general, would you say your health is [very good/good/fair/poor]?”; (2) “The quality of my life is [very good/good/fair/poor]?”; and (3) “Compared to one year ago, how would you rate your health now?” (with answer options “much better,” “somewhat better,” “about the same,” “somewhat worse,” or “much worse”). The questions have been administrated among elderly and Chinese populations.^33-35

Descriptive statistics were used to summarize sociodemographic characteristics, health status, quality of life, health change, and the prevalence of child maltreatment and intimate partner violence. In bivariate analyses, χ² tests were used to compare characteristics between individuals with or without child maltreatment and intimate partner violence experience. We also used χ² tests to compare experience of child maltreatment and intimate partner violence among individuals with or without elder abuse. To further determine the associations between child maltreatment and intimate partner violence, child maltreatment and elder abuse, and intimate partner violence and elder abuse, we computed multiple logistic regression models to control for potential confounding variables in basic sociodemographics, health status, quality of life, and health change. Five models were used with an increasing number of covariates. We included age and gender in model 1 and added education and income in model 2. In model 3, we further included living arrangement, marital status, and number of children. In model 4, we added culture-related variables, including years of living in the United States, country of origin, and language preference. In model 5, we added overall health status, health change over the past year, and quality of life. All statistical analyses were conducted using SAS, version 9.2 (SAS Institute Inc).

The prevalence of elder abuse, child maltreatment, and intimate partner violence was 15.2%, 11.4%, and 6.5%, respectively. Table 1 summarizes the sociodemographic information, health status, quality of life, and health change by the presence or absence of child maltreatment and intimate partner violence. As shown, among people who had child maltreatment and/or intimate partner violence experience, more were not from mainland China (12.3% vs 6.7%; 14.2% vs 6.91%). More people who indicated child maltreatment history were men (56.4% vs 40.2), whereas more people who indicated intimate partner violence history were women (65.7% vs 57.5%).

Table 2 summarizes the frequency of child maltreatment by intimate partner violence and elder abuse, as well as intimate partner violence by elder abuse. Intimate partner violence was more prevalent among people who reported child maltreatment (13.4% vs 5.6%). Elder abuse was more prevalent among those who reported child maltreatment (25.2% vs 13.8%) and intimate partner violence (48.8% vs 12.9%).

Results of the associations between child maltreatment, intimate partner violence, and elder abuse are summarized in Tables 3, 4, and 5. In the full models adjusting basic sociodemographic, health status, quality of life, and health change variables, people who reported child maltreatment had higher odds of intimate partner violence (adjusted odds ratio [aOR], 2.57; 95% CI, 1.78-3.71) and elder abuse (aOR, 2.08; 95% CI, 1.57-2.75). People who reported intimate partner violence had higher odds of elder abuse (aOR, 5.53; 95% CI, 4.01-7.64). English was grouped with Mandarin as a language preference for its small group size (n = 31); the results remained consistent after removing this group from the whole sample.

The present study examined the association between child maltreatment and intimate partner violence with elder abuse among a representative US Chinese older population. It appears to show that prior abuse is associated with higher odds of subsequent abuse, which extends the knowledge about interpersonal violence across the life course among older women and men.

Regardless of a number of sociodemographic, health status, quality of life, and health change covariates, this study indicates that individuals with a history of child maltreatment were associated with 2 times higher odds of intimate partner violence and elder abuse, and those who had experienced intimate partner violence were associated with almost 6 times higher odds of elder abuse. Although evidence is robust to support subsequent abuse after childhood sexual abuse, the findings of this study suggest a lasting association between early abuse on older adults.

The phenomenon of people experiencing repeated abuse has been conceptualized by a depth of prior sexual abuse and trauma research, which focused on individual-level factors that might expose individuals to higher risks of repeated abuse. For example, the learned helplessness concept suggests that those who have experienced abuse might develop chronic emotional distress, decreased motivation, and diminished confidence to strive against violence.^36,37 Future violence research should adopt more systematic frameworks, such as the life-course perspective,³⁸ to understand factors contributing to repeated abuse at the interpersonal, cultural, contextual, and societal levels.

Interestingly, being married was negatively associated with intimate partner violence, whereas there was no significant association between marital status and elder abuse. This contrast might indicate that elder abuse does not predominantly involve a partner—the major perpetrators in intimate partner violence. Indeed, a prior qualitative study shows that the Chinese older adults largely referred to elder abuse as in-family or intergenerational conflicts.³⁹ Additionally, more men reported child maltreatment whereas more women reported intimate partner violence. Resonant with the positive association between child maltreatment and elder abuse perpetration found in a previous study,⁴⁰ the findings of the present study might reflect a transmission of violence. Further investigations regarding abused-abuser dyads might add vital implication to the mechanism of the broader cycle of violence.

Findings of the relationship between education and elder abuse have been controversial,^41-43 and the results of this study demonstrate that higher levels of education are associated with higher odds of elder abuse. This might be related to people’s attitudes toward discussing adverse events with others; a previous study demonstrated that higher levels of education are associated with improved help-seeking attitudes toward mental health services.⁴⁴ The educational status may also be indicative of specific sociocultural characteristics that interact with the experience of violence, such as family structure and immigration experience. More mixed-method studies are needed for in-depth understandings of violence among minority ethnic and immigrant populations.

This study has limitations. First, the results might not be generalizable to other populations. Second, the data may be subject to recall bias because responses were based on retrospective memories. Third, given the face-to-face data collection format, abuse experience might be underestimated because of the social desirability bias. Fourth, there are mixed or incomplete psychometrical results for the accuracies of the Hwalek-Sengstok Elder Abuse Screening Test, Vulnerability to Abuse Screening Scale,^45-47 and Extended Hurt, Insult, Threaten, Scream scale, which need further examinations among US Chinese older adults. Fifth, the number of Chinese people who prefer to speak English is limited in this study, and future investigations are needed for this population. Lastly, there might be other unknown factors that have confounding effects on violence, such as sexuality, gender identity, and immigration status; LGBT+ and undocumented immigrant populations might be particularly vulnerable.

Despite the possible limitations, this study provides 2 important implications for health care and aging professionals. First, current evidence about screening elder abuse in primary care settings is inadequate,⁴⁸ and elder abuse might remain undetected during a brief clinical encounter. By understanding the interconnectedness of abuse across life span, health care and aging professionals could be more aware of the possibility of abuse when there is a known history of violence. Second, it is important to inform the health care and aging professionals that those experiencing elder abuse might have experienced other forms of violence, and the cumulative influence of violence on their health and well-being should be considered.

To further address gaps in knowledge, the interconnectedness of violence should be examined among large, nationally representative epidemiologic studies and other diverse populations. More importantly, given the lasting effect and incomplete evidence about best practices for violence,^48-52 rigorously designed randomized clinical trials are necessary for the development of effective prevention and intervention strategies for violence of different types. More psychometrical evaluations should be conducted to develop reliable and valid screening measures. As for policymakers, continuous legal assistance and other resources should be devoted to protecting the rights of individuals who are at higher risk of repeated abuse.

This study provides empirical evidence that people with experiences of child maltreatment and intimate partner violence are associated with higher odds of elder abuse. This continues to require leveraged and integrated efforts from researchers, practitioners, and policymakers to devise appropriate strategies for prevention of and intervention for people most likely to experience repeated abuse.

Accepted for Publication: January 27, 2019.

Corresponding Author: XinQi Dong, MD, MPH, Institute for Health, Health Care Policy and Aging Research, Rutgers University, 112 Paterson St, New Brunswick, NJ 08901 (xinqi.dong@rutgers.edu).

Published Online: May 20, 2019. doi:10.1001/jamainternmed.2019.0313

Author Contributions: Dr Dong had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: All authors.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: All authors.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Dong.

Obtained funding: Dong.

Administrative, technical, or material support: All authors.

Study supervision: Dong.

Conflict of Interest Disclosures: Dr Dong is supported by several grants from the Institute for Health, Health Care Policy and Aging Research at Rutgers University (P30AG059304, R01AG042318, R01MD006173, R01NR014846, R34MH100443, 90EJI0015, 90EJIG0016). No other disclosures are reported.